AI Won't Fix Your Discovery. But It Will Expose What's Missing

AI amplifies whatever process you already have

This is the part that doesn't get said enough about AI product discovery. If your discovery is structured, with clear questions, a systematic way of collecting evidence, and a method for connecting insights to decisions, then AI accelerates every step. Summarizing interviews, clustering feedback, surfacing patterns across channels. It's genuinely powerful.

If your discovery is ad hoc, AI gives you faster summaries of conversations you shouldn't have had, quicker synthesis of feedback you didn't collect systematically, and more polished artifacts that still don't connect to your roadmap decisions.

Teresa Torres, who created the Opportunity Solution Tree framework, has been making this point for years: the structure comes first. "You can't automate your way to good judgment. But you can automate the grunt work that prevents you from exercising judgment." The teams getting real value from AI discovery tools in 2026 are the ones who already had the scaffolding in place.

The gap is not tools. It's an operating model

Aha!'s 2026 release covers AI prompts for the full workflow from user interviews through feature specs. Productboard Spark is purpose-built for connecting customer insight to product decisions. These are real capabilities. But every PM tool I've looked at this year assumes the team already knows how to run discovery.

The teams I've seen struggle most are not missing an AI feature. They're missing the operating model: who talks to customers, how often, what questions get asked, where the insights go, and how they connect to what gets built. Without that, every AI tool is a point solution bolted onto a missing foundation.

My view is that the right order is: build the process first, even if it's simple. Get it working in a shared document. Then add AI to the steps that are repetitive. If you try to skip to AI-powered discovery without the underlying habits, you'll spend more time configuring tools than talking to customers.

What "18 steps" actually looks like

The Productboard data is worth unpacking. Going from 50 steps to 18 doesn't mean cutting corners. It means removing the manual labor between the steps that matter.

In practice, the teams I've seen do this well follow a pattern. They record customer conversations and let AI generate structured summaries against a consistent template. They use AI to cluster feedback from support tickets, sales calls, and NPS responses into opportunity areas. They still do the judgment work themselves: deciding which opportunities matter most, framing the problem clearly, and choosing what to test.

The 18 steps that remain are the high-judgment ones. The 32 that got removed were copy-paste, formatting, manual tagging, and synthesis work that a human was doing poorly because it was tedious.

That only works if you know which steps are high-judgment and which are grunt work. And you can only know that if you've done discovery manually enough times to tell the difference.

One thing to try before buying another tool

Pick your last three product decisions. Trace each one back to the customer evidence that informed it. If you can do that in under five minutes, your discovery process is strong enough to benefit from AI. If you can't, that's the thing to fix first.